PURPOSE: To determine whether magnetization transfer ratio(MTR) helps differentiate malignant from benignlesions of the head and neck. MATERIALS AND METHODS: In 36 patients with pathologically proven head and necklesions (malignant tumor, n=22; benign tumor and inflammation, n=14), we prospectively obtained axial SE T1, TSET2, and pre-and post-MT images(2D FLASH; TR/TE/=500/12, flip angle=20 degree) using a 1.5T superconducting unit. TheMT pulse used for MT images was 7 msec gaussian, with 2 kHz off-resonance. The signal intensities of ROI oflesions, muscle, fat, and CSF were measured during pre- and post-MT imaging. MTRs and corrected MTRs(cMTRs) werecalculated and compared between benign and malignant lesions. Statistical differences were evaluated by Wilcoxonrank sum and student t test. RESULTS: Statistically significant differences were found between MTRs and cMTRs ofmalignant and benign lesions(p<0.01) and muscle (p<0.01). The differences in MTRs and cMTRs of benign andfat(p<0.01) or CSF(p<0.01) were also statistically significant. In the case of malignant tumors, mean MTR and cMTRwere greater than those of benign lesions(0.343 +/-0.024 and 0.328 +/-0.026, vs 0.228 +/-0.049 and 0.193 +/-0.047,p<0.01). Using a criterion of 0.3 for malignancy, the diagnostic sensitivity, specificity, and accuracy of MTR formalignancy are 91, 93, and 92%, respectively. Using a criterion of 0.28, the corresponding figures for cMTR 95,93, and 94%, respectively. CONCLUSION: MTR analysis of MT imaging could help to differentiate malignant andbenign lesions of the head and neck.